CN104121949A - Condition monitoring method of ship electric propulsion system - Google Patents
Condition monitoring method of ship electric propulsion system Download PDFInfo
- Publication number
- CN104121949A CN104121949A CN201410404660.6A CN201410404660A CN104121949A CN 104121949 A CN104121949 A CN 104121949A CN 201410404660 A CN201410404660 A CN 201410404660A CN 104121949 A CN104121949 A CN 104121949A
- Authority
- CN
- China
- Prior art keywords
- data
- parameter
- sample
- electric propulsion
- neural network
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Abstract
The invention relates to a condition monitoring method of the ship electric propulsion system. The condition monitoring method is used for guiding the establishment of a condition evaluating system of the ship electric propulsion system, and belongs to the technical field of monitoring system design of the ship electric propulsion system. According to the method, on the basis of deep analysis of equipment operation principle of major components of the ship electric propulsion system, an intelligent evaluating algorithm based on knowledge is introduced in the condition evaluation of the ship electric propulsion system, the condition evaluation index system, procedure and algorithm of the ship electric propulsion system are firstly provided, and the condition evaluation model of the ship electric propulsion system and the condition evaluation steps of the ship electric propulsion system are established. The method has the advantages that the method has higher theory advancement and evaluation accuracy, and the method has guiding significance for practical application of the condition evaluation technology of the ship electric propulsion system.
Description
Technical field
The invention belongs to watercraft electric propulsion system Monitoring of Design Technology field, especially relate to a kind of watercraft electric propulsion system state monitoring method.
Background technology
Along with developing rapidly of electric semiconductor technology, AC speed-regulating theory and Control Technique of Microcomputer, watercraft electric propulsion system has had breakthrough progress at aspects such as maneuverability, reliability, operational efficiency, propeller powers.Its range of application constantly expands, and except being applied to the working ships such as ice-breaker, hog barge, ferry boat, is also widely used in the medium-and-large-sized conventional boats and ships such as oil tanker, pleasure boat, container ship, bulk freighter, has demonstrated wide market outlook.With its many superiority, oneself becomes the developing direction of Ship propulsion method to electric propulsion.Because the component devices of watercraft electric propulsion system is many, complex structure, and each equipment in service is to be mutually related, its safe condition is an integral body, so guarantee continuation, the reliability of watercraft electric propulsion system work, need to macroscopic view, to the security of operation of electric propulsion system equipment, carry out state estimation on the whole, find timely and fix a breakdown, reducing maintenance load, realizing State Maintenance and the robotization of electric propulsion equipment and control.
Summary of the invention
The present invention solves the existing technical matters of prior art; Provide a kind of and take watercraft electric propulsion system as object, set up the index system of electric propulsion system state estimation, there is higher theory advance and assessment accuracy, the practical application of boats and ships state assessment technology is directed concretely to the watercraft electric propulsion system state monitoring method of meaning.
Above-mentioned technical matters of the present invention is mainly solved by following technical proposals:
A watercraft electric propulsion system state monitoring method, following four index parameters of definition:
One group of reflection advances the index parameter of transformer state: comprise Static State Index parameter and dynamic indicator parameter;
Wherein, Static State Index parameter comprises technical parameter, test and overhaul data, same category of device fault history data; Described technical parameter comprises rated capacity, primary voltage, secondary voltage, efficiency and the class of insulation; Described test and overhaul data comprise each winding D.C. resistance, each winding insulation resistance, core inductance resistance, transformer aging conditions data and early stage operation troubles rate; Described same category of device fault history data comprise same category of device failure rate and same category of device most common failure data;
Wherein, dynamic indicator parameter comprises input parameter, output parameter, inner parameter and cooling system parameter; Described input parameter comprises original edge voltage, primary current and power input; Described output parameter comprises secondary voltage, secondary current and output power; Described inner parameter comprises each winding temperature of former limit, each winding temperature of secondary and Real time Efficiency; Described cooling system parameter comprises fan operation status data, inlet and outlet temperature, water pump operation status data, cooling water flow and intake-outlet temperature and pressure;
One group of reflection advances the index parameter of frequency converter state: comprise Static State Index parameter and dynamic indicator parameter;
Wherein, Static State Index parameter comprises technical parameter, test and overhaul data, same category of device fault history data; Described technical parameter comprises rated capacity, specified input voltage, rated frequency, output-current rating and efficiency; Described test and overhaul data comprise each device aging situation data, change device situation data and early stage operation troubles rate; Described same category of device fault history data comprise same category of device failure rate and same category of device most common failure;
Wherein, dynamic indicator parameter comprises input parameter, rectification parameter, inversion parameter, inner parameter and cooling system parameter; Described input parameter comprises input voltage, input current and power input; Described rectification parameter comprises each commutation diode operation conditions, DC bus-bar voltage, each device monitor data of filtering circuit and each device monitor of braking circuit; Described inversion parameter comprises each IGBT operation conditions, output voltage, output current, output frequency and output power; Described inner parameter comprise frequency converter from detection failure, frequency converter from detection alarm data, Real time Efficiency, internal temperature, interior humidity and well heater running status; Described cooling system parameter comprises fan operation status data, inlet and outlet temperature, water pump operation status data, cooling water flow and imports and exports cooling water temperature and pressure;
The index parameter of one group of reflection propulsion motor state: comprise Static State Index parameter and dynamic indicator parameter;
Wherein, Static State Index parameter comprises technical parameter, test and overhaul data, same category of device fault history data; Described technical parameter comprises rated power, rated voltage, rated frequency, rated speed, nominal torque, efficiency and the class of insulation; Described test and overhaul data comprise each winding insulation resistance, each winding D.C. resistance, motor aging conditions data and early stage operation troubles rate; Described same category of device fault history data comprise same category of device failure rate and same category of device most common failure;
Wherein, dynamic indicator parameter comprises input parameter, output parameter, inner parameter, cooling system parameter and bearing parameter; Described input parameter comprises input voltage, input current and power input; Described output parameter comprises current rotating speed, output power and load torque; Described inner parameter comprises each winding temperature of stator, speed setting, Real time Efficiency, interior humidity and well heater running status; Described cooling system parameter comprises fan operation state, inlet and outlet temperature, water pump operation state, cooling water flow and cooling water temperature and the pressure imported and exported; Described bearing parameter comprises each bearing temperature and transmission shaft vibration;
Comprise the following steps:
Step 1: the random following two kinds of methods of employing of selecting are monitored, and are respectively fuzzy neural network method and support vector machine method;
Wherein, in described fuzzy neural network method: definition N1～N4 is 4 one-level fuzzy neural networks, the sub-network that is called whole state estimation fuzzy neural network, 4 parts of expression watercraft electric propulsion system state estimation: advance transformer, advance frequency converter, propulsion motor and miscellaneous equipment; Wherein each sub-network is due to the difference of assessment apparatus, and input node and output node are also different, and nodal point number is by the training sample format determination of selecting; The residing state of equipment that the second layer of network model is judged according to ground floor is obtained the state estimation scoring of equipment
, the 3rd layer of basis
obtain the final scoring of watercraft electric propulsion system state estimation with the weight of each equipment
, comprise following sub-step: first definition has
individual state estimation index
,
plant fault
,
plant normal mode
equipment; Then:
Step 101, definite collection of passing judgment on
, set of factors
Pass judgment on collection
, set of factors
by training sample, determined,
,
;
Step 102, determine subordinate function: only have and determined fuzzy set
membership function, could set up fuzzy relation matrix
; Subordinate function adopts the terraced distribution of liter half in assignment technique and falls half terraced distribution, and function expression is:
formula one;
formula two;
Step 103, set up fuzzy relation matrix
with initialization neural network: simple fuzzy comprehensive evaluation method, the result of assessment depends on fuzzy relation matrix
accuracy, because subordinate function is determined difficulty, cause
poor accuracy, so affected the application of fuzzy comprehensive evoluation; And in fuzzy neural network, fuzzy relation matrix
just be used for the weights of initialization neural network,
on the not impact of the accuracy of trained neural network, and can improve the training speed of neural network and solve local minimum problem, so fuzzy neural network has an enormous advantage; Determined after subordinate function, just can set up fuzzy relation matrix
, then use
the weights of each connecting line of initialization neural network;
Step 104, sample normalization: refer to data parameters all in training sample to adopt extreme difference standardized method standardization, the data value after standardization exists
between, normalization sample can improve degree of accuracy, reduces the error of calculation, reduced data computing, the speed of convergence of quickening training pattern; Normalization formula is as follows:
formula three;
In formula,
the value for the treatment of normalizing in sample,
,
be respectively the very big and minimal value in this sample;
Step 105, neural network training: neural network is trained with the training sample after normalization; MATLAB simulation software provides Neural Network Toolbox, and tool box comprises the computing function of various neural networks, and wherein train (*) function is the training function of neural network, and this function can arrange training error and frequency of training, and calls conveniently;
Weights and the characteristic function of step 106, definite scoring: comprise two parts, the one, determine equipment judge collection
in respectively pass judgment on weights and the neuronic characteristic function of scoring of the line of rank and equipment state assessment scoring, the 2nd, determine weights and the neuronic characteristic function of overall score of the scoring of each equipment and the line of watercraft electric propulsion system state estimation overall score; Concrete, make a concrete analysis of according to the actual conditions of equipment and operation logic;
Step 107, calculating assessment result: complete after above-mentioned steps, just assessment data can be input in the fuzzy neural network of having trained, and calculate final scoring;
Wherein, described support vector machine method comprises following sub-step:
Step 111, determine support vector machine sorting algorithm, choose kernel function: the complexity based on sorting algorithm and the speed to sample training, generally choose one-to-many svm classifier or M-ary classification; Choosing of support vector machine kernel function has a significant impact the accuracy of test sample book assessment, in the problem of sample set being trained in concrete which type of kernel function of selection, because different samples, the diversity ratio of choosing different IPs function is larger, will be depending on sample so choose kernel function; Generally, gaussian radial basis function kernel function can reach requirement, is first-selected kernel function;
Determining of the normalization of step 112, sample and kernel functional parameter:
The method for normalizing of sample is identical with logical fuzzy neural network sample method for normalizing, refers to data parameters all in training sample to adopt extreme difference standardized method standardization, and the data value after standardization exists
between, normalization sample can improve degree of accuracy, reduces the error of calculation, reduced data computing, the speed of convergence of quickening training pattern; Normalization formula is as follows:
formula three;
In formula,
the value for the treatment of normalizing in sample,
,
be respectively the very big and minimal value in this sample;
The definite of kernel functional parameter is to obtain by the training sample set after normalization being searched for to study;
Step 113, to the study of training sample set and the assessment to test sample book: sample normalization and kernel functional parameter can be trained support vector machine with training sample after determining, then just can carry out state estimation to test sample book;
Step 2, input data: the following data constantly that comprise data, the historical data before current time and the prediction of watercraft electric propulsion system current time; According to the data of current time, can monitor the current operation conditions of watercraft electric propulsion system; According to the historical data before the data of current time and current time, and applied forcasting model, following data constantly can be obtained, thereby the following operation conditions constantly of watercraft electric propulsion system can be monitored, find potential fault and take measures in time;
Step 3, according to index system, electric power is carried out to system state monitoring to be divided into propelling Transformer's Condition Monitoring, to advance frequency converter status monitoring, propulsion motor status monitoring and state of other to monitor four parts, carry out respectively status monitoring, and then according to each several part, the weight of whole system is carried out to comprehensive monitoring to system, draw monitoring result, and by the running status of watercraft electric propulsion system be divided into excellent, good, in, poor, bad five class assessment results.
The application is creationary take watercraft electric propulsion system as object, set up the index system of electric propulsion system state estimation, take this index system carries out state estimation as basis, and the applied intelligent evaluation algorithm of this patent is different from the algorithm of the patent application occurring at present, has certain theory advanced
A kind of watercraft electric propulsion system state monitoring method above-mentioned, also comprises:
The index parameter of one group of reflection state of other monitoring index: comprise fault and warning message, each relay status data, each contactor status data, each fuse state data, generator operation status data, current supply voltage qualitative data and supervisory control system running situation data that current parking stall data, system operation troubles rate in early stage, PLC upload;
The external environment parameter of one group of reflection external environment situation: comprise ocean temperature, ship's heeling angle degree, boats and ships trim angle and ship hull vibration status data.
Therefore, tool of the present invention has the following advantages: have higher theory advance and assessment accuracy, the practical application of boats and ships state assessment technology is directed concretely to meaning.
Accompanying drawing explanation
Fig. 1 is watercraft electric propulsion system State Assessment Index System.
Fig. 2 is watercraft electric propulsion system state estimation flow process.
Fig. 3 is watercraft electric propulsion system fuzzy neural network state estimation model.
Fig. 4 is watercraft electric propulsion system support vector machine state estimation model.
Fig. 5 is fuzzy neural network state estimation step.
Fig. 6 is support vector machine state estimation step.
Fig. 7 a is for rising half terraced distribution plan.
Fig. 7 b is for falling half terraced distribution plan.
Embodiment
Below by embodiment, and by reference to the accompanying drawings, technical scheme of the present invention is described in further detail.
embodiment:
First applied statistics method is set up the index system of watercraft electric propulsion system state estimation, index is divided into and advances transformer index, propelling frequency converter index, propulsion motor index and miscellaneous equipment index, then propose the flow process of watercraft electric propulsion system state estimation and provide assessment result, watercraft electric propulsion system state estimation model and performing step based on intelligent evaluation algorithm are finally proposed, for the design of watercraft electric propulsion system status assessing system provides theoretical direction.
On technique scheme basis, further technical scheme is:
1, the foundation of watercraft electric propulsion system State Assessment Index System
Watercraft electric propulsion system mainly comprises and advances transformer, advances frequency converter, propulsion motor and miscellaneous equipment, and the state estimation of electric propulsion system is that to take the state estimation result of above-mentioned component devices be basis, through comprehensive assessment, obtains.The index system of watercraft electric propulsion system state estimation as shown in Figure 1, comprise and advance Transformer State Assessment index, propelling frequency converter state estimation index, propulsion motor state estimation index and state of other evaluation index, applied statistics method obtains index set in Table 1-table 4, in actual applications, can carry out suitable cutting according to the index set in actual conditions his-and-hers watches, thereby obtain being applicable to the State Assessment Index System of application.
Table 1 advances Transformer State Assessment index
Table 2 advances frequency converter state estimation index
Table 3 propulsion motor state estimation index
Table 4 state of other evaluation index
2, watercraft electric propulsion system state estimation flow process.
The fundamental purpose of watercraft electric propulsion system state estimation is to remove a hidden danger to take measures in time in order to assess the possibility of the current operation conditions of watercraft electric propulsion system and incipient fault generation; Will to watercraft electric propulsion system State Maintenance, propose guidance instruction according to the result of assessment on the other hand, aid decision making person formulates corresponding maintenance scheme according to the current state of watercraft electric propulsion system.Set up thus watercraft electric propulsion system state estimation flow process, as shown in Figure 2, watercraft electric propulsion system state estimation flow process is described below:
A. input data.
Input data are input state evaluation index, are the foundations of carrying out state estimation, input data, current time historical data before and the following data constantly of prediction that data comprise watercraft electric propulsion system current time here.According to the data of current time, can assess the current operation conditions of watercraft electric propulsion system; According to the historical data before the data of current time and current time, and applied forcasting model, following data constantly can be obtained, thereby the following operation conditions constantly of watercraft electric propulsion system can be assessed, find potential fault and take measures in time.
B. assessment algorithm is selected.
This patent adopts two kinds of algorithms: fuzzy neural network and support vector machine.Fuzzy neural network is the combination of fuzzy theory, artificial neural network, when possessing learning ability, can process nonlinear problem, is more satisfactory state estimation algorithm; Support vector machine method has stronger processing power to small sample data, is effective and practical state estimation algorithm.The state evaluating method that this patent relates to can select these two kinds of algorithms any one carry out state estimation.
C. watercraft electric propulsion system state estimation.
According to index system, electric power is carried out to system state assessment to be divided into propelling Transformer State Assessment, to advance frequency converter state estimation, propulsion motor state estimation and state of other to assess four parts, carry out respectively state estimation, and then according to each several part, the weight of whole system is carried out to comprehensive assessment to system, draw appraisal result.
D. assessment result is judged.
Centesimal system is taked in assessment result judgement, is 0～100 minute, within 0 minute, represents that state estimation index approaches or surpass the demand value of regulation; Within 100 minutes, represent that all state estimation indexs are all away from demand value or close with the factory-said value of quality product, electric propulsion is completely in normal condition.The condition grading of other situations is between 0 minute and 100 minutes.According to assessment mark, the running status of watercraft electric propulsion system be divided into excellent, good, in, poor, of inferior quality five class assessment results, in Table 5.
Table 5 watercraft electric propulsion system state estimation grade decision table
| Assessment mark | 81～100 | 61～80 | 41～60 | 21～40 | 0～20 |
| Evaluation grade | Excellent | Good | In | Poor | Bad |
E. assessment finishes.
Draw after assessment result and maintenance suggestion, judge whether to finish assessment, if otherwise continue input data and assess, otherwise assessment finishes.
3, watercraft electric propulsion system state estimation model.
One, fuzzy neural network state estimation model.
According to the ultimate principle of fuzzy neural network, in conjunction with the reality of watercraft electric propulsion system, three grades of fuzzy neural network models of design watercraft electric propulsion system state estimation, as shown in Figure 3.In accompanying drawing 3, N1～N4 is 4 one-level fuzzy neural networks, is called the sub-network of whole state estimation fuzzy neural network, represents 4 parts of watercraft electric propulsion system state estimation: advance transformer, advance frequency converter, propulsion motor and miscellaneous equipment.Wherein each sub-network is due to the difference of assessment apparatus, and input node and output node are also different, and nodal point number is by the training sample format determination of selecting.The residing state of equipment that the second layer of network model is judged according to ground floor is obtained the state estimation scoring of equipment
, the 3rd layer of basis
obtain the final scoring of watercraft electric propulsion system state estimation with the weight of each equipment
.
To have
individual state estimation index
,
plant fault
,
plant normal mode
equipment be example, the step of the state estimation based on fuzzy neural network is described.
A. determine and pass judgment on collection
, set of factors
.
Pass judgment on collection
, set of factors
by training sample, determined,
,
.
B. determine subordinate function.
Only have and determined fuzzy set
membership function, could set up fuzzy relation matrix
.Subordinate function has multiple definite methods such as expert's scoring, fuzzy statistical method, assignment technique, at present the most frequently used is the terraced distribution of liter half in assignment technique and falls half terraced distribution, as shown in Figure 7, wherein Fig. 7 (a) is for rising half terraced distribution, Fig. 7 (b) for falling half ladder distribution for its function.Function expression is:
(1)
(2)
In actual applications, for the be the bigger the better index of type of data, adopt to rise a half terraced distribution function, for the employing of the smaller the better type, half terraced distribution function falls, for those osculant data, get its absolute value and carry out data processing, if data absolute value increases progressively the employing that trend has been, rise half terraced function, adopt and fall half terraced function on the contrary.Parameter
,
generally according to the threshold value of index, determine.Determining of genus degree function also has some other method, be according to on-site actual situations and operating experience, and integrated application the whole bag of tricks.
C. set up fuzzy relation matrix
with initialization neural network.
Simple fuzzy comprehensive evaluation method, the result of assessment depends on fuzzy relation matrix
accuracy, because subordinate function is determined difficulty, cause
poor accuracy, so affected the application of fuzzy comprehensive evoluation.And in fuzzy neural network, fuzzy relation matrix
just be used for the weights of initialization neural network,
on the not impact of the accuracy of trained neural network, and can improve the training speed of neural network and solve local minimum problem, so fuzzy neural network has an enormous advantage.Determined after subordinate function, just can set up fuzzy relation matrix
, then use
the weights of each connecting line of initialization neural network.
D. sample normalization.
The normalization of sample refers to that the data value after standardization exists by data parameters employing extreme difference standardized method standardization all in training sample
between, normalization sample can improve degree of accuracy, reduces the error of calculation, reduced data computing, the speed of convergence of quickening training pattern.Normalization formula is as follows:
(3)
In formula,
the value for the treatment of normalizing in sample,
,
be respectively the very big and minimal value in this sample.
E. neural network training.
With the training sample after normalization, neural network is trained.MATLAB simulation software provides Neural Netword Toolbox (Neural Network Toolbox), tool box comprises the computing function of various neural networks, wherein train () function is the training function of neural network, this function can arrange training error and frequency of training, and calls conveniently.
F. determine weights and the characteristic function of scoring.
This part comprises two contents, the one, determine equipment judge collection
in respectively pass judgment on weights and the neuronic characteristic function of scoring of the line of rank and equipment state assessment scoring, the 2nd, determine weights and the neuronic characteristic function of overall score of the scoring of each equipment and the line of watercraft electric propulsion system state estimation overall score.Concrete, make a concrete analysis of according to the actual conditions of equipment and operation logic.
G. calculate assessment result
Complete after above-mentioned steps, just assessment data can be input in the fuzzy neural network of having trained, and calculate final scoring.
B. support vector machine state estimation model.
Two, the supporting vector machine model of watercraft electric propulsion system state estimation is as shown in Figure 4, similar to fuzzy neural network model, and just the state estimation sub-network of each equipment is different.Comprise the following steps:
A. determine support vector machine sorting algorithm, choose kernel function.
Complexity based on sorting algorithm and the speed to sample training, generally choose one-to-many svm classifier or M-ary classification.Choosing of support vector machine kernel function has a significant impact the accuracy of test sample book assessment, in the problem of sample set being trained in concrete which type of kernel function of selection, because different samples, the diversity ratio of choosing different IPs function is larger, will be depending on sample so choose kernel function.Generally, gaussian radial basis function kernel function can reach requirement, is first-selected kernel function.
B. the normalization of sample and kernel functional parameter determines.
The method for normalizing of sample is identical with logical fuzzy neural network sample method for normalizing, does not repeat.
The definite of kernel functional parameter is to obtain by the training sample set after normalization being searched for to study.
C. to the study of training sample set and the assessment to test sample book.
Sample normalization and kernel functional parameter can be trained support vector machine with training sample after determining, then just can carry out state estimation to test sample book.
Specific embodiment described herein is only to the explanation for example of the present invention's spirit.Those skilled in the art can make various modifications or supplement or adopt similar mode to substitute described specific embodiment, but can't depart from spirit of the present invention or surmount the defined scope of appended claims.
Claims (2)
1. a watercraft electric propulsion system state monitoring method, is characterized in that, following four index parameters of definition:
One group of reflection advances the index parameter of transformer state: comprise Static State Index parameter and dynamic indicator parameter;
Wherein, Static State Index parameter comprises technical parameter, test and overhaul data, same category of device fault history data; Described technical parameter comprises rated capacity, primary voltage, secondary voltage, efficiency and the class of insulation; Described test and overhaul data comprise each winding D.C. resistance, each winding insulation resistance, core inductance resistance, transformer aging conditions data and early stage operation troubles rate; Described same category of device fault history data comprise same category of device failure rate and same category of device most common failure data;
Wherein, dynamic indicator parameter comprises input parameter, output parameter, inner parameter and cooling system parameter; Described input parameter comprises original edge voltage, primary current and power input; Described output parameter comprises secondary voltage, secondary current and output power; Described inner parameter comprises each winding temperature of former limit, each winding temperature of secondary and Real time Efficiency; Described cooling system parameter comprises fan operation status data, inlet and outlet temperature, water pump operation status data, cooling water flow and intake-outlet temperature and pressure;
One group of reflection advances the index parameter of frequency converter state: comprise Static State Index parameter and dynamic indicator parameter;
Wherein, Static State Index parameter comprises technical parameter, test and overhaul data, same category of device fault history data; Described technical parameter comprises rated capacity, specified input voltage, rated frequency, output-current rating and efficiency; Described test and overhaul data comprise each device aging situation data, change device situation data and early stage operation troubles rate; Described same category of device fault history data comprise same category of device failure rate and same category of device most common failure;
Wherein, dynamic indicator parameter comprises input parameter, rectification parameter, inversion parameter, inner parameter and cooling system parameter; Described input parameter comprises input voltage, input current and power input; Described rectification parameter comprises each commutation diode operation conditions, DC bus-bar voltage, each device monitor data of filtering circuit and each device monitor of braking circuit; Described inversion parameter comprises each IGBT operation conditions, output voltage, output current, output frequency and output power; Described inner parameter comprise frequency converter from detection failure, frequency converter from detection alarm data, Real time Efficiency, internal temperature, interior humidity and well heater running status; Described cooling system parameter comprises fan operation status data, inlet and outlet temperature, water pump operation status data, cooling water flow and imports and exports cooling water temperature and pressure;
The index parameter of one group of reflection propulsion motor state: comprise Static State Index parameter and dynamic indicator parameter;
Wherein, Static State Index parameter comprises technical parameter, test and overhaul data, same category of device fault history data; Described technical parameter comprises rated power, rated voltage, rated frequency, rated speed, nominal torque, efficiency and the class of insulation; Described test and overhaul data comprise each winding insulation resistance, each winding D.C. resistance, motor aging conditions data and early stage operation troubles rate; Described same category of device fault history data comprise same category of device failure rate and same category of device most common failure;
Wherein, dynamic indicator parameter comprises input parameter, output parameter, inner parameter, cooling system parameter and bearing parameter; Described input parameter comprises input voltage, input current and power input; Described output parameter comprises current rotating speed, output power and load torque; Described inner parameter comprises each winding temperature of stator, speed setting, Real time Efficiency, interior humidity and well heater running status; Described cooling system parameter comprises fan operation state, inlet and outlet temperature, water pump operation state, cooling water flow and cooling water temperature and the pressure imported and exported; Described bearing parameter comprises each bearing temperature and transmission shaft vibration;
Comprise the following steps:
Step 1: the random following two kinds of methods of employing of selecting are monitored, and are respectively fuzzy neural network method and support vector machine method;
Wherein, in described fuzzy neural network method: definition N1～N4 is 4 one-level fuzzy neural networks, the sub-network that is called whole state estimation fuzzy neural network, 4 parts of expression watercraft electric propulsion system state estimation: advance transformer, advance frequency converter, propulsion motor and miscellaneous equipment; Wherein each sub-network is due to the difference of assessment apparatus, and input node and output node are also different, and nodal point number is by the training sample format determination of selecting; The residing state of equipment that the second layer of network model is judged according to ground floor is obtained the state estimation scoring of equipment
, the 3rd layer of basis
obtain the final scoring of watercraft electric propulsion system state estimation with the weight of each equipment
, comprise following sub-step: first definition has
individual state estimation index
,
plant fault
,
plant normal mode
equipment; Then:
Step 101, definite collection of passing judgment on
, set of factors
Pass judgment on collection
, set of factors
by training sample, determined,
,
;
Step 102, determine subordinate function: only have and determined fuzzy set
membership function, could set up fuzzy relation matrix
; Subordinate function has multiple definite methods such as expert's scoring, fuzzy statistical method, assignment technique, and at present the most frequently used is the terraced distribution of liter half in assignment technique and falls half terraced distribution, and function expression is:
formula one;
formula two;
Step 103, set up fuzzy relation matrix
with initialization neural network: simple fuzzy comprehensive evaluation method, the result of assessment depends on fuzzy relation matrix
accuracy, because subordinate function is determined difficulty, cause
poor accuracy, so affected the application of fuzzy comprehensive evoluation; And in fuzzy neural network, fuzzy relation matrix
just be used for the weights of initialization neural network,
on the not impact of the accuracy of trained neural network, and can improve the training speed of neural network and solve local minimum problem, so fuzzy neural network has an enormous advantage; Determined after subordinate function, just can set up fuzzy relation matrix
, then use
the weights of each connecting line of initialization neural network;
Step 104, sample normalization: refer to data parameters all in training sample to adopt extreme difference standardized method standardization, the data value after standardization exists
between, normalization sample can improve degree of accuracy, reduces the error of calculation, reduced data computing, the speed of convergence of quickening training pattern; Normalization formula is as follows:
formula three;
In formula,
the value for the treatment of normalizing in sample,
,
be respectively the very big and minimal value in this sample;
Step 105, neural network training: neural network is trained with the training sample after normalization; MATLAB simulation software provides Neural Network Toolbox, and tool box comprises the computing function of various neural networks, and wherein train (*) function is the training function of neural network, and this function can arrange training error and frequency of training, and calls conveniently;
Weights and the characteristic function of step 106, definite scoring: comprise two parts, the one, determine equipment judge collection
in respectively pass judgment on weights and the neuronic characteristic function of scoring of the line of rank and equipment state assessment scoring, the 2nd, determine weights and the neuronic characteristic function of overall score of the scoring of each equipment and the line of watercraft electric propulsion system state estimation overall score; Concrete, make a concrete analysis of according to the actual conditions of equipment and operation logic;
Step 107, calculating assessment result: complete after above-mentioned steps, just assessment data can be input in the fuzzy neural network of having trained, and calculate final scoring;
Wherein, described support vector machine method comprises following sub-step:
Step 111, determine support vector machine sorting algorithm, choose kernel function: the complexity based on sorting algorithm and the speed to sample training, generally choose one-to-many svm classifier or M-ary classification; Choosing of support vector machine kernel function has a significant impact the accuracy of test sample book assessment, in the problem of sample set being trained in concrete which type of kernel function of selection, because different samples, the diversity ratio of choosing different IPs function is larger, will be depending on sample so choose kernel function; Generally, gaussian radial basis function kernel function can reach requirement, is first-selected kernel function;
Determining of the normalization of step 112, sample and kernel functional parameter:
The method for normalizing of sample is identical with logical fuzzy neural network sample method for normalizing, refers to data parameters all in training sample to adopt extreme difference standardized method standardization, and the data value after standardization exists
between, normalization sample can improve degree of accuracy, reduces the error of calculation, reduced data computing, the speed of convergence of quickening training pattern; Normalization formula is as follows:
formula three;
In formula,
the value for the treatment of normalizing in sample,
,
be respectively the very big and minimal value in this sample;
The definite of kernel functional parameter is to obtain by the training sample set after normalization being searched for to study;
Step 113, to the study of training sample set and the assessment to test sample book: sample normalization and kernel functional parameter can be trained support vector machine with training sample after determining, then just can carry out state estimation to test sample book;
Step 2, input data: the following data constantly that comprise data, the historical data before current time and the prediction of watercraft electric propulsion system current time; According to the data of current time, can monitor the current operation conditions of watercraft electric propulsion system; According to the historical data before the data of current time and current time, and applied forcasting model, following data constantly can be obtained, thereby the following operation conditions constantly of watercraft electric propulsion system can be monitored, find potential fault and take measures in time;
Step 3, according to index system, electric power is carried out to system state monitoring to be divided into propelling Transformer's Condition Monitoring, to advance frequency converter status monitoring, propulsion motor status monitoring and state of other to monitor four parts, carry out respectively status monitoring, and then according to each several part, the weight of whole system is carried out to comprehensive monitoring to system, draw monitoring result, and by the running status of watercraft electric propulsion system be divided into excellent, good, in, poor, bad five class assessment results.
2. a kind of watercraft electric propulsion system state monitoring method according to claim 1, is characterized in that, also comprises:
The index parameter of one group of reflection state of other monitoring index: comprise fault and warning message, each relay status data, each contactor status data, each fuse state data, generator operation status data, current supply voltage qualitative data and supervisory control system running situation data that current parking stall data, system operation troubles rate in early stage, PLC upload;
The external environment parameter of one group of reflection external environment situation: comprise ocean temperature, ship's heeling angle degree, boats and ships trim angle and ship hull vibration status data.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410404660.6A CN104121949A (en) | 2014-08-18 | 2014-08-18 | Condition monitoring method of ship electric propulsion system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410404660.6A CN104121949A (en) | 2014-08-18 | 2014-08-18 | Condition monitoring method of ship electric propulsion system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN104121949A true CN104121949A (en) | 2014-10-29 |
Family
ID=51767457
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410404660.6A Pending CN104121949A (en) | 2014-08-18 | 2014-08-18 | Condition monitoring method of ship electric propulsion system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104121949A (en) |
Cited By (22)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105260786A (en) * | 2015-09-08 | 2016-01-20 | 哈尔滨工程大学 | Comprehensive optimization method of simulation credibility evaluation model of electric propulsion system |
| CN105741048A (en) * | 2016-02-23 | 2016-07-06 | 安徽容知日新信息技术有限公司 | Alarming method and apparatus for device |
| CN106355314A (en) * | 2016-08-26 | 2017-01-25 | 广东工业大学 | A Vehicle Engine Air Valve Quality Streaming Data Processing Method and Device |
| CN106408450A (en) * | 2016-09-09 | 2017-02-15 | 国家电网公司 | Power distribution capability evaluating method |
| CN106443275A (en) * | 2016-10-28 | 2017-02-22 | 中国舰船研究设计中心 | Method and apparatus for monitoring low-voltage distribution transformer for ship |
| CN106444597A (en) * | 2016-08-29 | 2017-02-22 | 苏州船用动力系统股份有限公司 | Power propulsion control system |
| CN107748557A (en) * | 2017-10-26 | 2018-03-02 | 哈尔滨工程大学 | A kind of Marine Synthesize Electric Propulsion System intelligent Fault Diagnose Systems based on LabVIEW |
| CN107769645A (en) * | 2016-08-22 | 2018-03-06 | 中车株洲电力机车研究所有限公司 | The cooperative control method of state estimation based on electric drive system |
| CN109145972A (en) * | 2018-08-09 | 2019-01-04 | 杭州电子科技大学 | A kind of watercraft electric propulsion system frequency converter alarm design method |
| CN109444575A (en) * | 2018-10-23 | 2019-03-08 | 长兴昇阳科技有限公司 | Motor driving system runs health state evaluation system and method |
| CN109747805A (en) * | 2018-12-28 | 2019-05-14 | 武汉船用电力推进装置研究所（中国船舶重工集团公司第七一二研究所） | A kind of energy-optimised management system of intelligent ship and method of hot redundancy |
| CN109978191A (en) * | 2019-03-29 | 2019-07-05 | 上海电气集团股份有限公司 | The appraisal procedure and assessment device of the system mode of industrial equipment system |
| CN110874088A (en) * | 2018-09-03 | 2020-03-10 | 中国船舶重工集团公司第七一一研究所 | Monitoring method of ship key equipment system based on multi-dimensional vector model |
| CN111301655A (en) * | 2020-03-18 | 2020-06-19 | 无锡赛思亿电气科技有限公司 | Ship electric propulsion and control system and monitoring control and safety evaluation method thereof |
| CN111445103A (en) * | 2020-02-25 | 2020-07-24 | 国网河南省电力公司电力科学研究院 | Power transmission cable production quality management feedback system based on industrial internet |
| CN111724289A (en) * | 2020-06-24 | 2020-09-29 | 山东建筑大学 | Environmental protection equipment identification method and system based on time sequence |
| CN112001028A (en) * | 2020-07-17 | 2020-11-27 | 武汉船用电力推进装置研究所（中国船舶重工集团公司第七一二研究所） | Simulation method and platform for permanent magnet propulsion system |
| CN112389607A (en) * | 2020-11-19 | 2021-02-23 | 哈尔滨工程大学 | Comprehensive power load prediction method for electric propulsion ship |
| CN113379005A (en) * | 2021-08-12 | 2021-09-10 | 新风光电子科技股份有限公司 | Intelligent energy management system and method for power grid power equipment |
| CN113433934A (en) * | 2021-04-27 | 2021-09-24 | 武汉海兰鲸科技有限公司 | Method for optimizing navigational speed of commercial ship |
| CN114863651A (en) * | 2021-01-15 | 2022-08-05 | 湖南五凌电力科技有限公司 | Intelligent early warning method for monitoring state of auxiliary machine |
| CN116361925A (en) * | 2023-05-31 | 2023-06-30 | 西北工业大学 | Multi-scheme assessment method and system for ship transmission configuration |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5713727A (en) * | 1993-12-09 | 1998-02-03 | Westinghouse Electric Corporation | Multi-stage pump powered by integral canned motors |
| JP2007132333A (en) * | 2005-11-13 | 2007-05-31 | Kimimasa Sumizaki | Water wave/wind force type water and land power generation plant |
-
2014
- 2014-08-18 CN CN201410404660.6A patent/CN104121949A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5713727A (en) * | 1993-12-09 | 1998-02-03 | Westinghouse Electric Corporation | Multi-stage pump powered by integral canned motors |
| JP2007132333A (en) * | 2005-11-13 | 2007-05-31 | Kimimasa Sumizaki | Water wave/wind force type water and land power generation plant |
Non-Patent Citations (3)
| Title |
|---|
| 何业兰等: "船舶电力推进系统状态评估软件设计与实现", 《武汉理工大学学报(交通科学与工程版)》 * |
| 王亚楠等: "基于分布式智能的船舶电力推进系统运行控制与管理策略研究", 《船电技术》 * |
| 王孟莲: "船舶电力推进系统状态评估研究", 《中国博士学位论文全文数据库 工程科技II辑》 * |
Cited By (29)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105260786B (en) * | 2015-09-08 | 2018-10-26 | 哈尔滨工程大学 | A kind of simulation credibility of electric propulsion system assessment models comprehensive optimization method |
| CN105260786A (en) * | 2015-09-08 | 2016-01-20 | 哈尔滨工程大学 | Comprehensive optimization method of simulation credibility evaluation model of electric propulsion system |
| CN105741048A (en) * | 2016-02-23 | 2016-07-06 | 安徽容知日新信息技术有限公司 | Alarming method and apparatus for device |
| CN107769645B (en) * | 2016-08-22 | 2020-04-21 | 中车株洲电力机车研究所有限公司 | Cooperative control method based on state estimation of electric drive system |
| CN107769645A (en) * | 2016-08-22 | 2018-03-06 | 中车株洲电力机车研究所有限公司 | The cooperative control method of state estimation based on electric drive system |
| CN106355314A (en) * | 2016-08-26 | 2017-01-25 | 广东工业大学 | A Vehicle Engine Air Valve Quality Streaming Data Processing Method and Device |
| CN106444597A (en) * | 2016-08-29 | 2017-02-22 | 苏州船用动力系统股份有限公司 | Power propulsion control system |
| CN106408450A (en) * | 2016-09-09 | 2017-02-15 | 国家电网公司 | Power distribution capability evaluating method |
| CN106443275A (en) * | 2016-10-28 | 2017-02-22 | 中国舰船研究设计中心 | Method and apparatus for monitoring low-voltage distribution transformer for ship |
| CN107748557A (en) * | 2017-10-26 | 2018-03-02 | 哈尔滨工程大学 | A kind of Marine Synthesize Electric Propulsion System intelligent Fault Diagnose Systems based on LabVIEW |
| CN107748557B (en) * | 2017-10-26 | 2020-05-15 | 哈尔滨工程大学 | Intelligent fault diagnosis system of ship comprehensive electric propulsion system based on LabVIEW |
| CN109145972B (en) * | 2018-08-09 | 2021-08-03 | 杭州电子科技大学 | Design method of frequency converter alarm of ship electric propulsion system |
| CN109145972A (en) * | 2018-08-09 | 2019-01-04 | 杭州电子科技大学 | A kind of watercraft electric propulsion system frequency converter alarm design method |
| CN110874088A (en) * | 2018-09-03 | 2020-03-10 | 中国船舶重工集团公司第七一一研究所 | Monitoring method of ship key equipment system based on multi-dimensional vector model |
| CN109444575A (en) * | 2018-10-23 | 2019-03-08 | 长兴昇阳科技有限公司 | Motor driving system runs health state evaluation system and method |
| CN109747805A (en) * | 2018-12-28 | 2019-05-14 | 武汉船用电力推进装置研究所（中国船舶重工集团公司第七一二研究所） | A kind of energy-optimised management system of intelligent ship and method of hot redundancy |
| CN109978191A (en) * | 2019-03-29 | 2019-07-05 | 上海电气集团股份有限公司 | The appraisal procedure and assessment device of the system mode of industrial equipment system |
| CN111445103A (en) * | 2020-02-25 | 2020-07-24 | 国网河南省电力公司电力科学研究院 | Power transmission cable production quality management feedback system based on industrial internet |
| CN111301655A (en) * | 2020-03-18 | 2020-06-19 | 无锡赛思亿电气科技有限公司 | Ship electric propulsion and control system and monitoring control and safety evaluation method thereof |
| CN111724289A (en) * | 2020-06-24 | 2020-09-29 | 山东建筑大学 | Environmental protection equipment identification method and system based on time sequence |
| CN112001028A (en) * | 2020-07-17 | 2020-11-27 | 武汉船用电力推进装置研究所（中国船舶重工集团公司第七一二研究所） | Simulation method and platform for permanent magnet propulsion system |
| CN112389607A (en) * | 2020-11-19 | 2021-02-23 | 哈尔滨工程大学 | Comprehensive power load prediction method for electric propulsion ship |
| CN112389607B (en) * | 2020-11-19 | 2021-10-01 | 哈尔滨工程大学 | Comprehensive power load prediction method for electric propulsion ship |
| CN114863651A (en) * | 2021-01-15 | 2022-08-05 | 湖南五凌电力科技有限公司 | Intelligent early warning method for monitoring state of auxiliary machine |
| CN113433934A (en) * | 2021-04-27 | 2021-09-24 | 武汉海兰鲸科技有限公司 | Method for optimizing navigational speed of commercial ship |
| CN113433934B (en) * | 2021-04-27 | 2022-04-12 | 武汉海兰鲸科技有限公司 | Method for optimizing navigational speed of commercial ship |
| CN113379005A (en) * | 2021-08-12 | 2021-09-10 | 新风光电子科技股份有限公司 | Intelligent energy management system and method for power grid power equipment |
| CN116361925A (en) * | 2023-05-31 | 2023-06-30 | 西北工业大学 | Multi-scheme assessment method and system for ship transmission configuration |
| CN116361925B (en) * | 2023-05-31 | 2023-11-03 | 西北工业大学 | Multi-scheme assessment method and system for ship transmission configuration |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104121949A (en) | Condition monitoring method of ship electric propulsion system | |
| CN108551167B (en) | XGboost algorithm-based power system transient stability discrimination method | |
| CN107941537B (en) | A kind of mechanical equipment health state evaluation method | |
| CN103091603B (en) | Breakdown intelligent classification and positioning method of electric transmission line | |
| CN106443447A (en) | An aero-generator fault feature extraction method based on iSDAE | |
| CN104976139B (en) | A kind of mechanical equipment state diagnostic method based on Gauss model | |
| CN106650122B (en) | A kind of equipment variable parameter operation methods of risk assessment | |
| CN110009208A (en) | A kind of on-pole switch complete set of equipments health state evaluation method and device based on integrated intelligent algorithm | |
| CN111190349A (en) | Method, system and medium for monitoring state and diagnosing fault of ship engine room equipment | |
| CN112200694A (en) | Dominant instability mode identification model construction and application method based on graph neural network | |
| CN110417005A (en) | In conjunction with the transient stability catastrophe failure screening technique of deep learning and simulation calculation | |
| Liu et al. | Deep learning for hardware-based real-time fault detection and localization of all electric ship MVDC power system | |
| Nasiri et al. | An open-water efficiency based speed change strategy with propeller lifespan enhancement in all-electric ships | |
| CN111950822B (en) | Performance detection method and device for vehicle traction transformer cooling system, storage medium and terminal | |
| CN106126949A (en) | A kind of steam turbine generator running status appraisal procedure | |
| Zarei et al. | Broken rotor bars detection via Park's vector approach based on ANFIS | |
| Voropai et al. | Preventive and emergency control of intelligent power systems | |
| Liu et al. | Dynamic security assessment of western Danish power system based on ensemble decision trees | |
| CN111814394B (en) | Power system safety assessment method based on correlation and redundancy detection | |
| Jose et al. | Fault diagnosis in induction motor using soft computing techniques | |
| Ma et al. | Fault diagnosis of fan gearboxes based on EEMD energy entropy and SOM neural networks | |
| Praveenkumar et al. | Intelligent fault diagnosis of synchromesh gearbox using fusion of vibration and acoustic emission signals for performance enhancement | |
| Skowron et al. | Application of hybrid neural network to detection of induction motor electrical faults | |
| Qi et al. | A Digital Twin Approach for Fault Diagnosis in Unmanned Ships Integrated Power System | |
| WANG et al. | Health Status Assessment Method of Oil Tanker Cargo Handling System |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| WD01 | Invention patent application deemed withdrawn after publication | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20141029 |